Vendor Status Watch

← all vendors

Anthropic status history

Current state: operational — “All Systems Operational” per https://status.anthropic.com (Atlassian Statuspage), checked 2026-09-16 12:28 UTC.

21incidents, last 30 days
54incidents, last 90 days
54incidents on record
Incidents per week, last 12 weeks:

How long do Anthropic incidents last?

61 minmedian incident duration54 measured
3.4 h90th percentile1 in 10 ran longer
86 minmedian, major/critical only20 incidents

From 54 incidents with a vendor-posted start and resolve time. Longest on record: Degraded functionality for Claude Cowork on Windows — 4.1 days (2026-09-10 15:54 UTC). Excluded, never estimated: nothing. Duration is the gap between the vendor's own posted timestamps — it is not measured downtime, uptime or an SLA figure, and a vendor that posts every blip will look worse here than one that posts nothing. Same rule as the outage-duration dataset.

Incidents on record (54)

StartedIncidentStatusImpactDuration
2026-09-15 09:48 UTCIntermittent error spikes for Claude Mythos 5.1 and Claude Fable 5.1resolvedmajor1 h 26 min
2026-09-11 14:19 UTCElevated errors for Claude Mythos 5.1 and Claude Fable 5.1resolvedmajor17 min
2026-09-10 21:43 UTCElevated latency on the Claude API for some usersresolvedminor1 h 40 min
2026-09-10 15:54 UTCDegraded functionality for Claude Cowork on Windowsresolvedmajor4 d 3 h
2026-09-03 13:26 UTCElevated errors for multiple modelsresolvedmajor2 h 57 min
2026-09-03 12:37 UTCElevated errors for Claude Sonnet 5resolvedmajor18 min
2026-09-02 21:17 UTCElevated errors for Claude Sonnet 5resolvedmajor26 min
2026-09-01 23:26 UTCDelays in credit purchasesresolvedminor1 h 57 min
2026-09-01 17:05 UTCDegraded performance on platform.claude.com and Claude for Microsoft Office 365resolvedminor1 h 01 min
2026-09-01 16:02 UTCDegraded performance on Claude for Microsoft Office 365resolvedminor19 min
2026-08-31 19:21 UTCDegraded performance on claude.ai and Claude Coderesolvedminor1 h 14 min
2026-08-31 17:23 UTCDegraded performance on claude.airesolvedminor29 min
2026-08-28 17:22 UTCElevated errors on Claude Code and Claude Coworkresolvedmajor2 h 59 min
2026-08-24 20:11 UTCIssues logging into Claude.airesolvedminor14 min
2026-08-24 17:12 UTCErrors logging into Claude.airesolvedminor3 min
2026-08-24 05:06 UTCElevated errors for multiple modelsresolvedmajor3 h 24 min
2026-08-20 19:16 UTCElevated errors on requests to multiple modelsresolvedmajor26 min
2026-08-20 18:32 UTCElevated errors on Google connectorsresolvedminor28 min
2026-08-19 09:42 UTCDegraded performance for Claude Opus 5 and Claude Haiku 4.5resolvedminor1 h 19 min
2026-08-18 16:20 UTCDegraded performance for multiple modelsresolvedminor2 h 41 min
2026-08-17 13:56 UTCDegraded performance for Claude Opus 5, Claude Sonnet 5resolvedminor1 h 32 min
2026-08-16 21:58 UTCService disruption on Claude servicesresolvedcritical35 min
2026-08-14 23:57 UTCElevated errors for Claude Fable 5resolvedmajor30 min
2026-08-14 20:30 UTCService disruption on Claude servicesresolvedmajor1 h 27 min
2026-08-14 07:58 UTCIssues reaching status.claude.comresolvednone2 h 55 min
2026-08-14 00:36 UTCService disruption on Claude Code on the web and related servicesresolvedminor14 min
2026-08-13 14:33 UTCElevated errors for Claude Mythos 5, Claude Fable 5, and Claude Sonnet 5resolvedmajor1 h 34 min
2026-08-12 13:50 UTCDegraded performance for multiple modelsresolvedminor4 h 16 min
2026-08-05 13:51 UTCDegraded performance for Claude Opus 5resolvedminor42 min
2026-08-05 07:05 UTCDegraded performance of multiple modelsresolvedminor7 h 08 min
2026-08-04 20:48 UTCElevated errors across many modelsresolvedmajor1 h 11 min
2026-08-04 11:27 UTCElevated errors on Claude Sonnet 5resolvedminor24 min
2026-08-04 06:16 UTCElevated errors on Claude Sonnet 5resolvedminor39 min
2026-08-03 15:13 UTCDegraded performance on Claude Sonnet 5resolvedminor15 min
2026-08-03 12:52 UTCError rates across multiple modelsresolvedminor1 h 24 min
2026-07-31 06:18 UTCDegraded performance on Claude Sonnet 5resolvedminor46 min
2026-07-30 13:43 UTCDegraded performance on Claude Opus 4.8resolvedminor40 min
2026-07-30 05:57 UTCElevated errors across many modelsresolvedmajor4 h 51 min
2026-07-29 19:49 UTCElevated errors across all modelsresolvedcritical2 h 46 min
2026-07-27 13:39 UTCElevated errors on Claude Opus 5 and Haiku 4.5resolvedminor56 min
2026-07-27 11:27 UTCElevated errors on Claude Opus 5resolvedminor1 h 02 min
2026-07-27 08:16 UTCElevated errors on Claude Opus 5resolvedminor48 min
2026-07-26 09:17 UTCElevated errors for Opus 5resolvedmajor1 h 27 min
2026-07-25 21:34 UTCElevated errors for Claude Fable 5, Claude Sonnet 5, Claude Haiku 4.5, and other modelsresolvedmajor33 min
2026-07-25 18:40 UTCElevated errors for Mythos 5, Fable 5, Opus 5 and Claude Haiku 4.5resolvedmajor1 h 04 min
2026-07-25 18:13 UTCSonnet 4.6 and Sonnet 5 errors elevatedresolvednone0 min
2026-07-24 09:09 UTCElevated error rate on Opus 4.8resolvedminor1 h 12 min
2026-07-22 23:52 UTCElevated errors on Haiku 4.5resolvedminor17 min
2026-07-22 17:56 UTCMicrosoft Office add-in availabilityresolvednone45 h 40 min
2026-07-22 17:27 UTCService disruption on Claude Servicesresolvedminor18 min
2026-07-22 15:24 UTCElevated error rates for Sonnet 5resolvedminor48 min
2026-07-22 13:37 UTCElevated errors on Opus 4.8 and Haiku 4.5resolvedminor1 h 46 min
2026-07-22 08:34 UTCElevated errors for Claude Opus 4.1resolvedminor25 min
2026-07-21 15:35 UTCElevated errors on several modelsresolvedcritical2 h 43 min

Titles, statuses and impact levels are the vendor's own words. Durations marked approx. are inferred from the poll that first found the incident gone. Shows the latest 60; the JSON has everything we hold (up to 400 days).

Watch Anthropic with the template

Put this in config.json of your fork and add a WEBHOOK_URL secret:

{
  "vendors": [
    "anthropic"
  ],
  "format": "auto",
  "map_url": "https://raw.githubusercontent.com/APVentureEngine/vendor-status-watch/main/vendors.json",
  "unreachable_after": 3,
  "mute_maintenance": false
}